Comorbidity Analysis for In-hospital Deaths++

Additional analyses for paid subscribers

Well, I just couldn’t resist poking this topic with GLM once more! In the previous post we were left with a statistically significant +0.19 diagnoses per electronic patient record for COVID in-hospital deaths in the post-vaccine rollout period compared to the pre-rollout period (p<0.001). My guess was this was due to age prioritised rollout and nothing more, so I did the obvious thing… restrict the sample to deaths for those aged 60 years and over in the post-rollout period and run GLM again using age as a covariate. Herewith the critical tables:

Our sample is now reduced to 1,588 COVID deaths in the post-rollout period of which 333 were vaccinated prior to death with either dose 1 alone or both doses (21.0%). In the bottom table we now see the Vaccination Status indicator variable now drop to p=0.104, rendering this statistically insignificant. Thus, when we account for age effects, even in a most basic fashion, we see no difference in the distribution of the diagnostic count between vaccinated and unvaccinated individuals; we may conclude that case complexity is equivalent.